Machine Learning Approach to Determine the Drug-Prone Areas in Lhokseumawe City, Indonesia

Abstract

This research aims to apply Machine Learning algorithms, such as K-means and K-nearest neighbor into a system that can determine drug-prone areas in Lhokseumawe City, Indonesia. The results of the k-means calculation in some clusters of areas that are very prone, prone and not prone to drugs based on drug user data with some criterias, namely the number of users, sub- district, education, occupation and types of drugs. Based on the research that tested by using the k-means algorithm on data on drug users in 2018-2021, the average percentage of clusters in very prone areas is 49.50%, the percentage of clusters in prone areas is 29.26% and the percentage of clusters in non-prone areas is 21.23%. Furthermore, the results of testing with the K-NN algorithm with 98 training data and 20 testing data obtained an average accuracy value of 87%.